The AI Pricing Showdown: OpenAI And Anthropic Compete For Subscribers

🔍 Read the full analysis: The AI Pricing Showdown: OpenAI And Anthropic Compete For Subscribers on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get the latest gadgets delivered free with Prime

• Fast, free delivery on millions of items
• Prime Video, Amazon Music and more included
• Member-only deals all year

Start your free Prime trial →Free trial for eligible customers · Cancel anytime

As an affiliate, we earn on qualifying purchases.Hide

OpenAI and Anthropic have cut prices for several AI models since June, and OpenAI is also offering discounts to attract customers from its rival, according to reporting by The Information. The changes are reshaping costs for businesses using AI, but headline savings depend on workload, contract terms and how much computing customers use.

Pricing figures compiled by ThorstenMeyerAI.com from vendor announcements and pricing pages show OpenAI’s Sol models moving from $5 input and $30 output per million tokens in June to $2 and $10 for GPT-6.1 Sol in September. Its smaller Luna model fell from $1 and $6 to 10 cents and 50 cents per million tokens. These are prices for successive models, not simply reductions to an unchanged model.

Anthropic also lowered prices. Its Opus line went from $5 input and $25 output per million tokens for Opus 5 in July to $4 and $20 for Opus 5.5. Anthropic separately reduced cache-read charges on some models. The company says those changes can lower costs on typical and agent-heavy workloads, but the savings vary with how a customer uses the models.

The Information reported on October 24 that OpenAI was offering discounts to attract enterprise customers. Reported examples include a discount of more than 30% for Replit on a contract valued at more than $100 million annually, according to a person cited in the report, and discounted seats for a US bank. DataCamp CEO Jonathan Cornelissen said he expected OpenAI’s pricing to cut his company’s costs by 80% to 90%. That figure is an expected saving against the company’s existing setup, not a published 90% reduction in OpenAI’s list price.

Lower Prices, More Contract Leverage

The cuts give companies more room to compare vendors and negotiate, particularly when their AI bills are large or they can move workloads between providers. Reported discounts add another layer to competition: the price on a public pricing page may not be the price a large customer receives. That makes contract terms and actual usage as important as per-token rates.

For buyers, a lower rate does not automatically mean a lower bill. Model performance, input and output volumes, cached-token charges, included usage and contract restrictions all affect total cost. Businesses also need to test whether a newer, cheaper model performs well on their own tasks rather than assuming that a successor is interchangeable with the model it replaces.

Individual subscribers face a related issue. ThorstenMeyerAI.com reports that OpenAI reduced usage included with new $200 Pro subscriptions while keeping the monthly price the same. A subscription can therefore become less generous even as model prices fall. Customers should check allowance limits and credit rules, not only the amount charged each month.

How the Pricing Cuts Accumulated

The reported changes came in a short sequence. OpenAI introduced lower-priced Luna models in July and cut Sol pricing in August, before announcing GPT-6 pricing in September. Anthropic set Sonnet 5’s introductory price as its standard price and later released Opus 5.5 at lower rates than Opus 5. The source’s pricing review says the listed figures match company announcements and current vendor pages retrieved on October 26.

Some savings come from cache pricing, rather than a reduction in the headline input and output rates. Caching can matter for AI agents that repeatedly use prior conversation context. Data scientist Max Woolf estimates cached tokens account for 96% to 99% of token spending in agentic coding; that is his estimate, not a figure independently established for every company’s workloads. Anthropic has said cache changes reduce costs on certain typical and highly agentic tasks.

The customer examples also have different levels of confirmation. Rippling CEO Parker Conrad said OpenAI’s share of the company’s coding work rose from 16% in May to 52%, while Anthropic’s fell from 57% to 18%, citing cost as the reason. Replit CEO Amjad Masad said the company had a preferred partnership with OpenAI, while also telling TechCrunch that Replit remained model agnostic. A person familiar with the partnership told TechCrunch it was still in place.

“I think ChatGPT is better now.”

— Amjad Masad, Replit CEO, speaking at TechCrunch Disrupt

Discounts and Usage Terms Vary

The available reporting does not provide complete contracts for the customers cited, so the exact discount conditions, duration and usage commitments are not clear. The reported bank arrangement and Replit discount are attributed to sources familiar with the deals, rather than full public contract disclosures. It is also unclear how broadly OpenAI is offering comparable terms.

The 80% to 90% estimate from DataCamp is a company-specific expectation. It should not be read as a guaranteed saving for other customers or as a direct comparison of list prices. Actual costs will depend on workloads and the mix of models used. Nor do lower prices establish that successor models match older models on every task; the source notes that customers need to test them against their own requirements.

Some plan changes add uncertainty for subscribers and enterprise buyers. The source reports that Anthropic added capped, expiring API credits to Max and Team plans and requires a 20-seat minimum for Enterprise. It also says OpenAI reduced usage included with new Pro subscriptions. The full effects depend on each plan’s current terms and a customer’s usage, which can change over time.

Watch Renewals and Plan Limits

Businesses considering renewal or a vendor change can compare current written offers, test models on representative tasks and calculate total cost across token use, caching, seats and included credits. Routing work across more than one provider may give buyers flexibility, but the source does not establish that any particular customer will benefit from doing so.

Further evidence will come from updated pricing pages, contract renewals and customer disclosures showing whether the reported discounts persist. Subscribers should review usage allowances and expiration rules alongside monthly fees. It remains to be seen whether the price competition continues at the same pace or shifts toward private enterprise concessions and changes to what plans include.

Key Questions

Did OpenAI cut its prices by 90%?

No general 90% list-price cut is reported. The 80% to 90% figure came from DataCamp’s CEO describing expected savings against his company’s existing setup. Reported OpenAI discounts include a deal of more than 30% for Replit.

How much did OpenAI’s model prices fall?

According to pricing figures checked by ThorstenMeyerAI.com, GPT-6.1 Sol was listed at $2 per million input tokens and $10 per million output tokens in September, compared with $5 and $30 for the earlier Sol model in June. Luna prices also fell, but the comparisons are between different model versions.

Has Replit stopped using Anthropic?

The available accounts do not establish that Replit has stopped using Anthropic. CEO Amjad Masad said OpenAI was Replit’s preferred partner and also said the company remained model agnostic. A person familiar with the partnership told TechCrunch it remained in place.

Why do cache prices matter?

AI agents may repeatedly process prior conversation context. Lower cache-read prices can reduce costs for those workloads even when a model’s headline input and output prices stay the same. The size of the saving depends on the customer’s usage.

What should AI subscribers check?

Check included usage, seat minimums, credit caps and expiration dates as well as the monthly fee. The source reports changes to usage in OpenAI’s new $200 Pro subscriptions and to credits and seat requirements on some Anthropic plans.

Source: ThorstenMeyerAI.com